Learning (and Teaching) Latin Verb Tenses: Applying Second Language Acquisition Research and Analyses of Verb Uses in Context

(1)

Learning (and Teaching) Latin Verb Tenses:

Applying Second Language Acquisition Research and Analyses of Verb Uses in Context

¹

Rebecca Harrison Truman State University

A

^bstrAct

This article uses research from Second Language Acquisition and from analyses of tense and verb uses in Latin authors to shed light on how students acquire verb tenses, especially past tenses, and to show how to enhance student learning of Latin verb tenses. I argue that 1) only one tense and one translation/use of a tense should be taught at a time; 2) the simple past meaning of the perfect should be the first past tense taught, then the stative use of imperfect, followed by the past progressive use, the present perfect, and pluperfect; 3) it makes a difference what verbs are used in teaching a tense; there are different semantic categories of verbs, such as stative, based on their inherent meaning, and tenses are acquired best with the category of verb that best matches the tense aspect; the combination of category of verb and tense can also sometimes aid in distinguishing translations of tenses; and 4) tenses occur in typical semantic and syntactic contexts, and these contexts are important to enhance acquisition and to aid in distinguishing different tense uses/translations.

Finally, I provide teaching suggestions, sample exercises, methods for adapting and supplementing textbooks, and a review of textbooks in relation to teaching verb tenses.

K

^ey

W

^ords

Verb tenses in Latin, perfect tense, imperfect tense, pluperfect tense, Caesar, Vergil, lexical aspect

Many of today’s Latin textbooks present the imperfect as the first past tense, often at the same time as the future, and the perfect tense after the imperfect, sometimes together with the pluperfect and future perfect. They also often teach all the translations/uses for a tense at once. Some textbooks use the same one verb as the

1 Thanks to my readers and colleagues, including Amy Norgard, Greg Richter, Tom Capuano, Masa- hiro Hara, James Hammerstrand and Tim Farley for sharing their expertise and/or for their comments and advice, and especially to John Gruber-Miller for his suggestions. Thanks also to Keely Lake for all her work converting formatting. Any remaining errors are mine.

Harrison, Rebecca. “Learning (and Teaching) Latin Verb Tenses: Applying Second Language Acquisition Research and Analyses of Verb Uses in Context.” Teaching Classical Languages 10.2: 1-47. ISSN 2160-2220.

(2)

standard in paradigms and in presenting new tenses. These practices seem to have little theoretical basis other than tradition or an assumed ease or efficiency in learning.

In this paper, I argue that only one tense and only one translation/use of a tense should be taught at a time. Further, the first past tense that should be taught is the simple past use of the perfect. After this should be the stative use of the imperfect, which regularly expresses an on-going physical, mental, or sometimes emo- tional state with (stative) verbs such as sum, iaceō, putō, or amō, then the past progressive use of the imperfect, followed by the present perfect use of the perfect, and finally the pluperfect. Second, I argue that the choice of verbs used in examples in teaching a new tense can aid or impede acquisition; tenses are acquired best with the type of verb whose inherent meaning best matches the tense aspect, such as currō, an activity, for the progressive use of imperfect. Third, the acquisition of verb tenses within characteristic contexts, such as with associated adverbs, and often certain other tenses, or in certain syntactic constructions, such as a relative clause, etc. is also important in acquiring tenses and developing appropriate expectations for reading.

These arguments rely on two strands of research. First, I reviewed the research findings from Second Language Acquisition (SLA)² related to acquiring tenses, especially past tenses, to see how the findings can apply to and enhance teaching Latin verb tenses. In the first part of this article, I present some basic findings and summarize five underlying principles from SLA. Second, I created my own corpus analysis of tense and verb uses in four Latin authors. My research is based on an analysis of uses of indicative tense verbs from a sample of Latin texts, including book 1 of Caesar’s De Bello Gallico, Cicero’s De Amicitia, books 1, 2, and 4 of Vergil’s Aeneid, and Plautus’ Mostellaria. I chose these as a representative sample of different genres and types from authors that students are likely to read. I also used Oldsjö’s analyses of tense uses in Caesar. The emphasis in this study is on indicatives, usually the first forms taught.

These two strands of research provide the empirical evidence for the three major topics in this article.

1. The order of acquisition of past tenses and the factors enhancing or impeding acquisition.

2 I use the term SLA in this article to refer to research studies on acquisition by immersion or by instruction or in combination.

(3)

2. The categorization of verbs into four semantic types based on their inherent meaning, and the relationship between the semantic category of verb and tense aspect. The combination of type of verb and tense can enhance (or impede) student learning of a tense and can also sometimes aid in distinguishing uses of tenses and translations of verbs and in translating indirect statement.

3. The characteristic contexts for each past tense in Latin so that teach- ers can provide authentic, quality input for instructional examples and exercises, and so that students can learn to predict and interpret the verbs met in reading.

After discussing the characteristic context results, I then go through each tense in Latin in the recommended order (simple past use of the perfect, stative imperfect, past progressive use of the imperfect, present perfect use of the perfect, and pluperfect) with suggestions for teaching, including sample exercises. Finally, I provide suggestions for adapting and supplementing textbooks and a review of textbooks in relation to teaching verb tenses and aspect.

I will begin with the summary of some basic background findings from SLA research.

s

^econd

L

^AnguAge

A

^cquisition

r

^eseArch

: b

^AcKground

Developing an understanding of the temporal system of a second language takes time. It is more than the sum of the individual parts, and it evolves gradually and is revised as more tenses and observations are added (Bardovi-Harlig, Tense 95). The temporal system includes temporal markers, such as adverbs, as well as verb morphology. There are also limits on the memory and processing ability of language learners. Below are summaries of five underlying principles that will recur in discussions of one or more of the three major topics of this article: 1) stages of temporal acquisition, 2) the One to One Principle, 3) processability and Pienemann’s teachability hypothesis, 4) the distinction between explicit and implicit knowledge, and 5) the importance of quality input in instruction. I will then discuss applications of SLA research to the common practices of teaching 1) multiple translations/uses of a tense and 2) multiple tenses at the same time as a preliminary to my first major topic, the order of acquisition of past tenses.

(4)

1) Stages in Temporal Acquisition: SLA research has shown that when peo- ple learn a language, whether by immersion or instruction 1) they rely first on the order in the narrative for temporality, i.e. assuming chronological order; then, 2) they use lexical clues from adverbs, prepositional phrases, etc.; and 3) finally, the morphological forms are learned and used (Ortega 126; Bardovi-Harlig, Tense 111, 414-15, 420-21; Ellis and Sagarra 590).

2) One to One Principle: Andersen’s study of learners’ limited processing ability and concept development shows that as second language learners try to con- struct a consistent system, they initially follow what he termed the One to One Prin- ciple: one form equals one translation/use, that is, learners need a one to one correspondence as they are first learning (Andersen 77-95).

3) Processability and Pienemann’s Teachability Hypothesis: Some features of language, such as vocabulary, are variational in that they do not have to be learned in a certain order. Other things are developmental and are acquired in a certain order (Lightbown and Spada 177-78). For these developmental features, Pienemann has argued that there are certain stages and sequences of learning; learners must learn certain concepts before they are ready to learn the next (or a later) stage in the sequence. This is the Teachability Hypothesis.

4) Explicit vs. Implicit Knowledge: There is a difference between explicit knowledge and implicit learning or understanding and how they are used. It is the difference between what one may know vs. what one can do. For example, I had students come into intermediate Latin who could identify the tense of a verb form (in isolation) or complete a paradigm (explicit knowledge), but could not translate the verbs correctly in context (implicit understanding). Explicit and implicit knowledge are dissociated in separate parts of the brain and do not always interface (Cintrón-Valentín and Ellis 201). As VanPatten has argued, it is implicit knowledge that is used in reading and comprehension; explicit instruction is not necessarily ap- plied, though it may facilitate learning (VanPatten 29-35, 58-59) when it is provided in teaching a tense that is ready by processability theory to be learned. See more below in the discussion of Interaction with Input, Salience, L1 Influence and the Role of Explicit Instruction in Temporal Acquisition.

5) The Importance of Quality Input in Instruction: The quality of the input and intensity of interaction are factors of critical importance that can enhance the speed of (or impede and delay) acquisition (Ortega 141-42, Bardovi-Harlig, Tense 432-33). Authentic or appropriate input is the basis for the development (over time) of the full implicit understanding of each tense and the distinction between that tense and others (VanPatten 29-35, 58-59). Components of input that will be discussed include the frequency of occurrences, the salience (how easy it is to notice the form)

(5)

and interaction with input, the semantic aspect of the verbs themselves and the relationship between the verb semantic type and tense aspect, and the various features of context including such things as associations with certain words, syntactic constructions, and other tenses.

Applications to Common Teaching Practices

1) Argument against Teaching Multiple Translations/Uses at the Same Time - One to One Principle and Processability:

Does teaching all the translations/uses of a tense at the same time enhance acquisition? This practice goes against the One to One Principle, which shows the need for one tense form - one translation at first in the development of the temporal system (Ortega 127; Bardovi-Harlig, Tense 425-26). While exposure to the concept that a form does not necessarily have one equivalent translation is important, and a preview of the different uses can give a big picture view, which can benefit certain types of learners, learners initially establish one meaning/translation, which should be the basic, most common one, as their default in working memory. For the Latin perfect, this basic translation is the simple past, which is the first past tense acquired in the order of acquisition, as discussed below; by processability the present perfect use should not be taught until later, after the imperfect and past progressive. Learn- ing a language is like getting directions to a place you have not been. One initial set of directions for the basic route is useful; variations can be confusing at first. With experience, alternatives for special situations, such as rush hour congestion or rain, make more sense and can be processed and remembered more easily.

2) Arguments against Teaching Multiple Tenses at the Same Time:

SLA research also shows that it is best to teach only one tense at a time. One reason is that teaching complete tense systems, such as the perfect, pluperfect, and future perfect together, or the imperfect and future together, can be confusing because of memory overload and confusion of similar forms. Words and forms that are similar in phonetic/orthographic form are more apt to be confused with each other, especially if they are taught at the same time (Sommers and Lewis 83-108; Laufer 153). Differences that are only a different vowel in the same syllable position in words are especially apt to cause confusion (Laufer 146-48). Confusion (by cross association) is also apt to occur when the words are similar or opposites in semantic idea (Schmitt 147; Nation 47). Both problems apply to teaching the imperfect and future at the same time: non-present (essentially similar in being “opposites” of

(6)

present and of each other); new tenses marked by a morpheme beginning with “b” in the same syllable differing only by the vowel. Similarly, the pluperfect and perfect tense are both past tenses, with the third person plurals similar except for one vowel and the quantity (and thus the accent): ērunt vs. erant. The future perfect, another non-present tense, is also similar to the pluperfect, except for the vowel in almost all the forms.

So, if possible, it is better to teach the imperfect and the future tenses separately, and, similarly, the perfect, pluperfect and future perfect separately.³ Nation suggests that presentation of such words or forms should occur weeks to months apart to allow sufficient time after the first form is firmly settled before introducing the second similar word or form; the time needed varies for individuals (Nation 47).

If one needs to teach multiple tenses with similar orthography, such as imperfect and future, at the same time or closely together, using illustrative pictures with cap- tions can help reduce the confusion (Schacter, Israel, and Racine 1-24). Mnemonic methods do not work as well with phonologically or orthographically similar words unless the keyword uses a distinguishing characteristic (Hulstijn 203-24). I have used as examples: for the imperfect, -ba- = wa(s) (___ing), vs. future, -bi- = wi(ll), and “Will Bo or Bill bunt the ball to advance the runner from first to second base [conjugations]?”; and for the pluperfect, “If the ERA had passed, women . . .”

A more fundamental reason for not teaching multiple tenses simultaneously is that, by processability theory, teaching the pluperfect together with the perfect does not follow the order in which past tenses are acquired. Textbooks that present the tenses in batches to get through the material as quickly as possible and get to reading real Latin inhibit their goal by their method, which can result in possible confusion and explicit rather than implicit learning, which is needed for processing and reading. The instructional emphasis should be on what students can do (implicit knowledge) in reading, rather than on what they know (explicit knowledge),

3 Insufficient research has been done for the acquisition of the future tense relative to the different past tenses. The future is one of the basic tenses, in Donatus (see below under Suggestions for Instruction) and in SLA research. It is used more frequently in speech than in narrative. The future perfect makes up only about 1% of all verb tense forms, and also occurs most often in subordinate clauses, especially conditions and temporal clauses (Mahoney 102; Haverling 381). Since English does not use a future in subordinate clauses, the most common translation in English is, in fact, not the actual future perfect “will have ___ed,” but a “present” or perfective “have ___ed” (see also Wigtil 682, 685). Thus, the future perfect is best learned with or after conditions.

(7)

commonly emphasized in grammar textbooks and traditional testing.⁴ The most ef- fective instruction follows the order that students need in order to learn, with ad- equate time for acquisition, which is my first major topic.

o

^rder^of

A

^cquisition^of

P

^Ast

t

^enses

The order of acquisition of past tenses is a developmental feature that follows the Teachability Hypothesis. SLA research has shown that the simple past is the first past tense acquired, with the stative use of the imperfect next, appearing much later (Ortega 127-29; Bardovi-Harlig, Tense 114, 227, 419-22; Andersen and Shirai 559). After these come: 3) past progressive tense (if it is present in the language) or use of the imperfect (with the iterative and habitual developing later); 4) present perfect; and then 5) pluperfect (Bardovi-Harlig, Tense 115, 117, 419-23, 429; Ortega 127). SLA research also shows the following points: 1) This order of acquisition of past tenses of verbs is consistent across languages, including the learning of English and other Germanic languages (German, Dutch, and Swedish), and the Romance languages, including French, the closest to Latin in terms of tenses, as discussed below. 2) This order is consistent whether acquisition is by immersion or by instruction or a combination. 3) The order of acquisition is not affected by the order in which the tenses are taught; instruction can change the rate of learning, but not the order of acquisition of tenses (Bardovi-Harlig, Tense 111-12, 405-06). In particular, learners have been found to acquire and show a real understanding of the concept of the simple past form first, even when other tenses have been introduced (Bardovi-Harlig, Tense 111-12). Bardovi-Harlig attributes this consistent order of acquisition to the order of development of concepts based on Pienemann’s teachability hypothesis (Tense 392-95), which explains why it holds even when tenses are taught in a different order.

Factors Affecting the Rate of Acquisition

There are also factors that SLA research has identified that can influence the rate of acquisition, either to enhance and speed up learning or delay or impede learning a tense (Ortega 140-42, Bardovi-Harlig, Tense 432-33, Lightbown and Spada 48). These include: 1) the complexity or clarity of the form; 2) the frequency; 3) the quality and intensity of interaction with input; 4) the salience of input; and 5)

4 See my article “Exercises for Developing Prediction Skills in Reading Latin Sentences,” TCL 2.1 (Fall 2010): 1-30.

(8)

the similarity or difference between the first (L1) and second language. I will first discuss the complexity of form and then the frequency. I will show in the third major section on characteristic contexts that teaching the simple past use of the perfect before the imperfect in Latin also enhances acquisition because of the salience of the unmarked use of the simple past, and the regular use of the imperfect in context with the perfect (discussion with Table 6), meaning the imperfect cannot be used in appropriate, authentic input contexts without the perfect. In addition, as discussed below in the Suggestions for Instruction, the lack of a separate stative imperfect form in English L1 can cause confusion when the imperfect is taught first before the simple past use of the perfect because of the translations, e.g. “was,” often being the same.

1) Ease of Form: One common assumption for teaching the imperfect as the first past tense in Latin is the “ease” of the form. However, the “ease” of the form is not the primary consideration in learning tenses, and, in fact, irregular forms of basic tenses are learned before regular ones. Easier forms can speed up learning the forms themselves, but they do not affect the order of the acquisition of tense concepts (Bardovi-Harlig, Tense 112, 420-21). Research in French and Spanish, where the preterite morphology is more difficult or complicated (involving stem changes, etc.) than the imperfect, similar to Latin, shows that the simple past is still acquired first before the imperfect (Bardovi-Harlig, “The Place” 28-29). The simple past, the first past tense acquired, is actually the most often inflected, whereas other tenses may involve paraphrases with auxiliary verbs or similar (Andersen and Shirai 561), reflecting their less basic nature.

Thus, the assumption that the imperfect should be taught first before the perfect in Latin because the imperfect form is easier in that the morphology is consistent and new stem(s) do not have to be learned is not supported by SLA research.

Learning the morphological form is not the same as learning the tense concept, and the development of the tense system is implicit, not explicit knowledge (Bardovi- Harlig, Tense 112, 421; VanPatten 22, 42, 58-59). Students can memorize forms and complete paradigms (explicit learning), but if that tense is not what is needed for the first or next tense in the development of the temporal system, the knowledge will not be integrated and used in processing or reading. In particular, students can give/“translate”/identify forms of the imperfect and do exercises using explicit knowledge, or even imitate without really understanding, but they cannot truly acquire the imperfect until they have learned the simple past (Fraser, Bellugi, and Brown 121-35).

(9)

2) Frequency: A second factor that can speed or delay learning a tense is the frequency in input. The perfect is the most frequently used past tense according to Oldsjö’s study of Caesar (using his data for indicative tenses) and is corroborated by my study of indicative verbs in Caesar, Cicero, Plautus, and Vergil. I also include here in Table 1 Mahoney’s comprehensive analysis of tense frequency use in the larger corpus of Latin literature in general, which, however, includes all moods (including participles and infinitives, etc.), but excludes irregular verb forms (Mahoney 102).

Present Perfect^a Imperfect Pluperfect Future

Overall (including all moods, but not irregular forms) (Mahoney 102)

54.0% 29.6% 7% 2.2% 6.3%

(+ Future perfect:

0.9%) Caesar BG

(Oldsjö 310) 940^b (= 28%) (+ 200 ambiguous present/perfect = 6%)

955 (= 29%)^c (+ 200 ambiguous present/perfect);

(890 simple past

= 27%; 65 present perfect = 2%)

774 (=23%) 463

(= 14%) NA

Caesar BG 1 147 = 26%

(+ 34 ambiguous present/perfect = 6%)

196 = 34%

(+ 34 ambiguous present/perfect) (193 simple past

131= 23% 66 = 11% 0^d

Cicero De Am.

1.1-9.32 204 = 62% 79 = 24%

(54 simple past = 16%; 25 present perfect = 8%)

24 = 7%

(48 total in the whole De Am.)

7 = 2%

(10 total in the whole De Am.)

13 = 4%

(+ 2 future perfect) Plautus Most.

(lines 1-531) 355 = 63% 116 = 21%

(78 simple past = 14%; 38 present perfect = 7%)

9 = 2%

(19 total in the whole work)

5 = 1%

(10 total in the whole work)

69 = 12%

(+ 10 future perfect) Vergil Aen.

book 1 355 = 58%

(+ 4 ambiguous present/perfect = 1%)

137 = 22%

(+ 4 ambiguous present/perfect) (120 simple past

45 = 7%

(134 total in books 1, 2, and 4)

17 = 3%

(42 total in books 1, 2, and 4)

51 = 8%

(+ 2 future perfect)

Table 1: Relative Frequency of Tenses in Latin

(10)

a. As in Table 4 and Table 7, data for the perfect does not include all of Cicero’s De Amiticia, Plautus’

Mostellaria, and books 2 and 4 of Vergil’s Aeneid because of the sheer number of perfects and the impracticality of using search functions to check that all are accounted for. For the imperfect (Tables 5 and 6) and pluperfect (Table 8), the numbers are totals for the whole work or book(s) indicated. I also include the total figures for imperfect and pluperfect in this table in part to indicate how much more frequently the perfect tense is used.

b. Oldsjö only gives figures for the historic present use, not all present tense verbs. He also does not include the future.

c. For main clauses, which is what learners would be seeing most initially, Oldsjö also gives figures for indicative (and historic infinitive) uses for five different historical authors, with Caesar (BG &

BC) the lowest at 33.7% perfect (compared with 21.4% imperfect and 5.6% pluperfect) compared to 61.3% perfect for Livy, 63% for Florus, 79.5% for Velleius Paterculus, and a high of 94.3% for Eutropius (278-80).

d. There were no future indicatives in my analysis. There were future infinitives (and subjunctives) in indirect statement.

Even with the present perfect uses of the perfect accounted for (see the table above), the simple past use of the perfect is still used more frequently than the imperfect, which is, in turn, used more frequently than the pluperfect. Thus, the order by frequency corresponds in general to the order of acquisition of past tenses in SLA research and supports teaching the perfect tense before the imperfect in Latin and teaching the simple past use of the perfect first before the present perfect. The different uses for each tense will be discussed more below.

3-5) Interaction with Input, Salience, L1 Influence, and the Role of Explicit Instruction in Temporal Acquisition:

Salience, another factor in the rate of acquisition, is important in getting learners to notice and acquire information. If learners can get away without learning something by relying on other features, they will (Ortega 126; Bardovi-Harlig, Tense 111, 414-15, 420-21; Ellis and Sagarra 590). As indicated above, the use of chronological order and lexical adverbs, etc. precede the third stage of acquisition, the use of the verb forms themselves. In order to acquire the verb tense form – meaning associations and gain implicit knowledge, learners must pay attention to the verbs.

Ellis and Sagarra’s study shows that this preference for lexical temporal cues, such as herī, over morphological form, such as cogitāvī, applies to the learning of Latin. The study also showed that there is influence from the L1, which, as indicated, is another factor that can enhance or impede acquisition. English L1 learners showed more reliance on temporal adverbs and less on verb morphology than Rus- sian and especially Spanish L1 learners, who had experience with more similar verb morphology in their L1.

(11)

Cintrón-Valentín and Ellis then performed a follow-up study to ascertain which method of explicit instruction and/or implicit exposure to input was most ef- fective in overcoming the attention bias to lexical cues and the blocking effects of the use of such cues on learning verb morphology in Latin. The research was based on an experiment with three activities, using three temporal adverbs, hodie, herī, cras, and the first, second, and third person singular forms of cogitō in the present, perfect, and future. There was also monitoring of eye-tracking. Their results showed that the control group (with no training other than the feedback, “correct” or “incor- rect” plus the correct temporal reference, during the first activity, which involved identifying the temporal reference of adverb-verb/verb-adverb pairs as past, present, or future), as predicted, relied on the lexical adverbs and did not direct sustained attention to or learn the verb forms. Two other groups showed increased reliance on and knowledge of the verb morphology, but with a lower rate of learning the lexical adverbs as shown in subsequent activities, which the authors attributed to limited working memory in processing.⁵ The first of these groups had an introduc- tory grammar lesson of four slides (without practice) on the formation of the verb morphology, including the use of blue color for “tense markers” and red for (subject)

“agreement markers.”⁶ The second group had attention directed to the morphemes through highlighting of the inflection with bold and red font during the first activity.

The most balanced learning of both the lexical vocabulary and verb morphology (for all three tenses including the future) happened in the third group. This group first had a short multiple choice pre-training session of feedback (“correct” or “in- correct” plus the correct translation) on the present and perfect (but not future) verb forms and their translation (for each of the six verb forms six times in random order) (Cintrón-Valentín and Ellis 203-04, 207-10, 217-18, 223, and 229-32).

Cintrón-Valentín and Ellis’ follow-up study confirmed that explicit instruction with practice, especially for formations that are different in the second language than the first, must come before exposure to meaningful input to allow the learners

5 In the second activity, they were given the individual adverbs, individual verbs, and adverb-verb/

verb-adverb pairs, which could be matched or mis-matched in time. They were again asked to identify the temporal reference (past, present, both past and present, etc.), but without feedback. This was followed by the third activity asking for written production of translations from English to Latin of just adverbs, just verbs, and adverb-verb sentences (e.g. “Yesterday he thought.”) without feedback.

(Cintrón-Valentín and Ellis 209-10).

6 I have used similar color highlighting when teaching verb tenses, but I use different colors for the tense markers and subject endings. I use blue for the subject endings, since I also use blue for the nominative case endings. I use green or yellow for the imperfect/future tense inflection marker, and pink for perfect (active) stems. For example, currēbam, vocāvit.

(12)

to integrate the processing of both the new forms and new lexical vocabulary (229- 32). Thus, presenting the forms and their meaning with practice can enhance learning, if done when the learner is ready to acquire that tense. Knowledge of both the verb forms and context clues, including lexical vocabulary, is important, especially in Latin, as shown below in the third major section on characteristic contexts.

M

^Atching^of

t

^ense

A

^sPect^And

V

^erbAL

s

^eMAntic

A

^sPect

(q

^uALity

i

^nPut ^And

o

^ne^to

o

^ne

P

^rinciPLe

)

When teaching the verb forms and their associated tense meaning, an important component of the quality of input that has been found to enhance (or delay) acquisition of tenses is the choice of verbs in input, in particular, matching the type of verb used with each tense, which is my second major topic. This is the Lexical Aspect Hypothesis developed by Andersen and Shirai, based on the work of Vendler and extended by others. They distinguish two kinds of aspect: 1) One is based on the grammatical verb tense, including the traditional punctual or aorist, (the simple past use of the Latin perfect), the imperfective, the progressive, and the perfective (with a resulting state) (Latin present perfect use of the perfect). 2) The other kind of aspect is that related to the semantic meaning of the verb itself, the inherent lexical aspect.

The four basic semantic categories of verb types originally identified by Vendler, based on whether the inherent verb meaning: expresses punctuality or duration; has a set end point (telic) or not (atelic); and is dynamic (requires energy) or not, are shown in Table 2 (Andersen and Shirai 531-32; Bardovi-Harlig, Tense 193, 214-23, 425):⁷ For each category, I include examples of verbs in Latin that are frequently used and common in textbooks that would be a good match in teaching.

Semantic Category

of Verb Type Characteristics Prototypical

Tense Match Common Latin Verbs

Achievement punctual (instantaneous event as a whole); telic;

dynamic

simple past inveniō; relinquō;

dō; mittō;

(ad)veniō; vocō;

reperiō; ponō Accomplishment durative; telic;

dynamic; (completion of a process or action with set beginning and end and an outcome)

present perfect interficiō; claudō;

faciō; occīdō;

discō; līberō;

intrō; parō

7 For a five category situation aspect, including more sentence components, see Bachvarova 124-25.

See also Oldsjö 161-65, 170-75, 227, 439; Haverling 288-310; and Miller 18-19, 23-24.

(13)

Activity durative; atelic;

dynamic progressive (use of

imperfect) dūcō; currō;

scrībō; cōgitō;

eō; audiō; legō (“read”); dīcō;

teneō; portō; trahō Stative durative; atelic; not

dynamic stative (use of)

imperfect sum; possum;

habeō; vīvō; putō;

timeō; iaceō;

sedeō; taceō; amō;

crēdō Table 2: Lexical Aspects of Verbs

Both achievements and accomplishments have an inherent end point, or are

“telic,” and can be used with an ablative of time, but not an accusative of duration.

Achievements express an instantaneous event as a whole and cannot be used with

“stop” or “continue” doing it, while accomplishments express more the completion of a process (or action) with a set beginning and end (and an outcome) (Bardovi- Harlig, Tense 215-16; Haverling 305; Andersen and Shirai 531-32; Ortega 127).

Verbs of activity, which match most closely with a progressive tense, express a homogeneous process which requires continuing energy. Unlike accomplishments and achievements, they have no set end point and they may be used with an accusative duration of time (Bardovi-Harlig 215-16, 223; Andersen and Shirai 532).

Stative verbs inherently persist over time without continuing energy and without a set end point. Like verbs of activity, they can occur with an accusative duration of time, but unlike activity verbs, they do not tend to occur in the impera- tive mood, and they are not used in the progressive tense (in languages that have a separate progressive tense) or a progressive use of the imperfect tense. If a state ceases, a different new state becomes (Bardovi-Harlig, Tense 215-16, 223). Note that included in the stative verbs in the table above, some verbs of emotion and mental verbs may express states.⁸ Second conjugation verbs are also especially common as intransitive states.

SLA research, especially by Andersen and Shirai, on first and second language acquisition, including classroom instruction and multiple languages, has shown that tenses are acquired first and best with the semantic category of verb that matches most closely with the aspect of the particular verb tense. The lexical aspect type that matches also has the highest frequency with that tense in authentic

8 For more on mental states, see Bardovi-Harlig 219 and Oldsjö 438-39.

(14)

input, since the correspondence of categorization of verbs with tense also reflects the normal distribution patterns of verbs by tense in the language. This “prototypical”

matching is related to the One to One principle: each tense is initially associated with the one lexical category it (cognitively) matches most closely (Andersen and Shirai 548-55). Thus, the match of verb type and tense is important in the examples and sentences/passages when teaching. This means that when presenting a new tense, one should choose example verbs (from the verbs that students have had or are currently learning) from the category that best matches that tense in semantic aspect, such as mīsit (achievement) for the simple past use of the perfect, in order to enhance student learning.

Verb Types in Indirect Statement: Note that for indirect statement, the con- cept of categories of verbs is also useful in understanding when English can “fudge”

the tenses of infinitives with secondary main verbs and when not. For example, for perfect infinitives, “He said that he sent (mīsisse) it” (achievement), will work instead of the more literal “that he had sent it,” but for vīxisse (stative), “He said that she lived” will not work for “that she had lived.” Conversely, for present infinitives, (mittere) “He said that he sent it” will not work for “that he was sending it,” but for vīvere, either “He said that she lived” or “that she was alive/living” will work.

Extension of Tenses by Verb Types: After learners acquire a tense with the prototypical type of verb, they then may extend their understanding of the tense through the acquisition of other lexical categories of verbs in order, as shown in the table below:

Simple Past: Achievement Accomplishment Activity Stative Imperfect: Stative Activity Accomplishment Achievement Progressive: Activity Accomplishment Achievement (xx: no stative) Present

Perfect: Accomplishment Achievement (activity/stative)^a Table 3: Acquisition of Lexical Aspect Types by Tense

Sources: Andersen and Shirai 529-30, 555-59; Ortega 114, 127-28; Bardovi-Harlig, Tense 94, 227, 422-30, 434.

a. The sources do not give an order for the extension to verbs of activity vs. statives.

Note that the order of acquisition is the opposite for the simple past and the imperfect. This reflects the correlation between the natures of the verbs and the tenses. For teaching new tenses, using examples of verbs of the type that are learned last (or not used) for that tense do not aid in learning, such as the stative amāvit for the simple past. Thus, the same verb that the textbook uses as the standard in its paradigm may not fit well for the particular new tense or new translation being

(15)

taught. In general, verbs of activity, such as dūcō, match best with the most tenses for a paradigm.

As indicated, after verbs of achievement, the simple past is extended in acquisition to verbs of accomplishment and later of activity, and gradually to stative as the temporal system is developed and refined based on input (Bardovi-Harlig 423;

429; Andersen and Shirai 559). The present perfect is associated most closely, as indicated, with verbs of accomplishment (and achievement) (Ortega 127; Andersen and Shirai 559). Since both uses of the Latin perfect - simple past and present perfect - occur most often with the same two categories of verbs, accomplishment and achievement, the context is the best guide in translation, as discussed below under context uses (Table 7).

The use of the secondary extended verb types often include contextual mark- ings and differences in meaning or tense use. For the imperfect, as indicated, the prototypical match is the stative, then activities (Ortega 114, 128; Bardovi-Harlig, Tense 94, 227, 422, 424, 426, 429, 434; Andersen and Shirai 559). The extension of the imperfect to verbs of activity corresponds initially and most commonly to the progressive use of the imperfect, especially for languages like Latin that did not add a separate progressive tense formation, and it follows the order of acquisition of tenses: first stative imperfect, then progressive. Verbs expressing an activity, which match most closely with the progressive, are the most frequently used type of verb in the progressive use of imperfect in Latin in my analysis. The category of verb, in fact, functions as a guide in the most likely choice of translation of the imperfect:

activity as progressive, stative as stative imperfect. When the simple past translation is included for the imperfect in textbooks, it should be noted that it is used mainly for stative verbs and is not the most likely translation for other types of verbs.

Extended Uses of Verbs, Verb Types, and Tense Uses: Another way of look- ing at the extension of verb types used in tenses is that verb types may sometimes be extended to other tenses that they are not the prototypical match for. These less frequent extended uses are often distinguished by context and/or meaning/tense use, or by voice (as videō/videor below). There are often differences when verbs of activity, which match most closely with the progressive (use of imperfect), are used in the perfect. For example, in my analysis, when currō occurs in the perfect, there is usually a prepositional phrase, “per ____,” giving an end point, and the use is often metaphorical with an abstract subject (e.g. tremor, fremor, calor) (e.g. Vergil Aen.

2.120-21).

(16)

Another example of such extended use is that of achievement and accomplishment verbs, which match most closely with the simple past use of the perfect and the present perfect. In the imperfect, such extended use tends to express iteratives (cf. Oldsjö 227), the repetition of an action by an individual or multiple individuals, or habitual actions. For example, the only example I found of inveniō in the imperfect (compared to five perfects in Caesar, two in Cicero, one in Plautus, and nine in Vergil) was iterative (Caesar BG 2.16.1 cum . . ., inveniēbat ex captīvīs plus an indirect statement).⁹ Thus, the type of verb is an important indicator of the iterative or habitual translation for the imperfect.

Verbs can also sometimes belong to more than one category, depending on the meaning or narrative context use (situation aspect), e.g. with a direct object vs.

intransitive, or whether the object is concrete or abstract (Bachvarova 125; Oldsjö 151, 165-70).¹⁰ Videō is an example of a common, but more complex verb. In my analysis, in the perfect it regularly expresses its basic meaning of physical sight;

the uses of videō in the imperfect are regularly mental perception (stative), usually with an indirect statement (or an abstract object, such as ventura in Aen. 2.125), or occasionally iterative (e.g. Vergil Aen. 8.360 marked by passim: passimque armenta vidēbant . . . mugīre), or are most often passive (meaning “seemed,” 75% of the im- perfects in Cicero, 48% in Caesar, and 40% in Vergil), and are then stative.

Conclusion: The extension of the concept of tenses to include these other semantic types of verbs and uses are part of the gradual expansion and refinement of the development of the temporal system and also sometimes of the vocabulary of the verbs themselves, such as videō, and their varied meanings and uses. Such extension is acquired by implicit learning from input when the learner is ready for the next stage; instruction can aid acquisition by providing appropriate input contexts and/or directing attention to them as needed, such as by asking leading questions, which is often more appropriate when students encounter them in reading actual texts. It is like learning to predict a storm based on previous observations of seeing the clouds getting dark, the wind picking up, and hearing rumbling in the distance or seeing lightning.

This variability emphasizes the need for quality input in terms of appropriate contextual usages in examples, exercises, and readings. As argued above with the One to One principle, initial input with the most common use in their characteristic

9 Similarly, the only occurrence of the simple verb mittō in the imperfect in all of Caesar BG, Cicero De Am., Plautus Most., and Vergil Aen., is an iterative in BG 5.45.1 (Quanto . . . tanto crebriores lit- terae nuntiique ad Caesarem mittebantur).

10 For more on the historical development, complexities, and/or studies on a particular author, see, for example, Oldsjö (on Caesar and as compared to other Latin historians) and Haverling.

(17)

context is important: verbs of achievement, such as inveniō, for the simple past use of the perfect; stative verbs, such as sum, for the imperfect; verbs of activity, such as legō, for the progressive use of the imperfect; and verbs of accomplishment, such as interficiō, for the present perfect use of the perfect. Less common secondary uses, such as the iterative use of the imperfect, or meanings need to be in their characteristic contexts in order for learners to integrate these later and to develop appropriate expectations. As shown, knowledge of the semantic types of verbs is sometimes useful in distinguishing different uses of tenses in translating. The flex- ibility in verb meanings and tense uses in Latin may be related to the relatively late position of the verb in Latin clauses, and the role of context in distinguishing these uses also highlights the importance in Latin of reading in order, rather than hunting for the verb first, as is sometimes recommended. Jumping to the verb bypasses the clues from the gathering storm clouds, etc.

c

hArActeristic

c

^ontexts

My third major topic is what the characteristic contexts are in which each of the different tenses most commonly occur and their relationship to tense uses and semantic verb types. The contexts support the order of acquisition of tenses and can help a reader anticipate what tense a verb will be and/or, as indicated, interpret which use of a tense or the translation of a verb may be. They are based on the findings from my analysis of Latin authors and that of Oldsjö on Caesar. Table 4 below shows the context uses of the simple past verbs, the first tense in the order of acquisition. The context uses of the present perfect use of the perfect will be discussed below (Table 7), after the imperfect (Tables 5 and 6), and lastly the pluperfect (Table 8), following the order of acquisition of tenses.

Simple Past Use of the Perfect

Compared to the other past tenses, the simple past use of the perfect, as the basic past tense, does not generally need or have characteristic context markers, making the verb more salient.

(18)

Caesar BG & BC (Oldsjö 310, 321, 331, 333, 339)

Caesar BG 1 Cicero De Am. 1.1- 9.32

Plautus Most.

lines 1-531

Vergil Aen.

book 1 Main Clause 1436 = 89%

(+ 335 ambiguous present/perfects)

176 = 91%

26 = 48% 54 = 69% 92 = 77%

Subordinate

clause 178 = 11%

17 = 9%

(+ 5 ambiguous present/perfect)

28 = 52% 24 = 31% 28 = 23%

Relative 50 regular

+ 4 other relative 4 (+1 ambiguous

present/perfect) 19 10 8

Temporal 90 11 (including

9 ubi) (+ 3 ambiguous present/perfect)

3 8 17

Other 38 2 (+1 ambiguous

present/perfect) 6 6 3

Table 4: Contexts of Simple Past Use of Perfect Indicative Verbs

For use in subordinate clauses, the frequency of relative clauses reflects the general frequency of relative clauses in Latin, with qui/quis being second on Diederich’s list of Latin words by frequency (Diederich 115). The relatively higher percentage of subordinate clauses in general for the simple past use of the perfect in Cicero compared to the other authors reflects the relative complexity of his style.

Of the 97 occurrences of simple past verbs in a subordinate clause in my analysis of Caesar, Cicero, Plautus, and Vergil, only five of these verbs (in four sentences), had an imperfect verb in the main clause (erat in three sentences and videbatur in one).

The imperfect, by comparison, as shown below, is regularly used in context with a perfect tense verb, which supports teaching the simple past use of the perfect before the imperfect.

Imperfect Tense Uses and Their Characteristic Contexts

As indicated above, the imperfect is the second most frequently used past tense in Latin and includes the prototypical stative and then progressive use, which are the second and third past tenses acquired in SLA. As shown in Table 5 below, the stative is the most frequent use of the Latin imperfect in my analysis of Caesar, Cicero, and Plautus, and second most frequent for Vergil. The habitual and the

(19)

iterative uses tend to occur in clusters, so the data by clause rather than by number of verbs would probably reflect a smaller percentage of uses for the habitual/

iterative use consistent with the order of frequency and SLA order of acquisition.

Cicero’s philosophical dialogue, De Amicitia, also probably has a higher proportion of habitual/iteratives than his speeches due to the characters, setting, and discussion content. The progressive use is the least common (by number of verbs) in Caesar, Cicero, and Plautus, probably due to Caesar’s more informative style and Cicero’s more philosophical style; it is the most common in Vergil and may be related to batching and to his higher percentage of uses in main clauses and his poetic tone of narrative.

Caesar BG 1 Cicero de Am. Plautus Most. Vergil Aen.

books 1, 2, and 4 Stative Imperfect 100 = 76% 24 = 50% 9 = 47% 38 = 28%

Progressive use 9 = 7% 6 = 12.5% 2 = 11% 81 = 60%

Habitual/iterative use 22 = 17% 18 = 37.5% 8 = 42% 15 = 11%

Table 5: Uses of the Imperfect Indicative Verb Forms

In all four authors the verb sum is the most frequently used verb in the imperfect (twenty-four times = 18% in Caesar BG 1, ten times = 21% in Cicero, six = 32% in Plautus, and thirteen = 10% in Vergil), consistent with the relatively high frequency of the stative imperfect use. The next most frequently used verbs in Caesar are also stative: possum, compounds of sum, habeō, and mental verbs of thinking.

Cicero also has mostly stative (and habitual activity, as shown below) verbs after sum, with three times each for habeō, putō, videor (and videō one time) and dīcō.

Vergil has teneō, dō, ferō, trahō, and eō for his next most frequently used verbs in the imperfect, reflecting his higher percentage of progressive uses with verbs of activity.

My analysis shows how the context and/or verb type may aid in anticipating an imperfect tense and/or distinguishing between these different uses/translations of the imperfect. Some uses and/or contexts are also more common to certain authors and/or genres.

(20)

Caesar BG & BC (Oldsjö 310-11, 315-16, 331-39)

Caesar BG 1 Cicero de

Am. Plautus

Most. Vergil Aen.

books 1, 2, and 4

Main Clause: 1,013=

64% 75 = 56.5% 25 = 52% 14 = 74% 106 = 79%

causal: Nam(que)/enim 6 1 0 7

Iam(que)/tum/temporal adv. 1 4 3 25

demonstrative/connecting

relative 27 6 3 9

hīc/place connector 0 0 1 6

Verb of saying (usually iterative/habitual) or thinking, etc. + ind.

statement

20 (+3:

oportebat, volebat, habebat + acc. + inf.)

4 (+1:

volebat + acc. + inf.)

4 (aieba[n]t) 2

(especially progressive) with pluperfect in same (or coordinate) clause

2 0 0 9

summary talis, etc. 0 0 0 15

adversative autem/immo/at/

sed/vero/tamen 1 6 1 1

autem (moreover)/etiam 6 1 0 0

existential sum (+ dative(s)) 4 (1) 1 (0) 0 (2) 4 (2)

With inverse cum clause 0 0 0 12^a

ecce 0 0 0 4

Other^b 15 7 2

(habitual) 23 Subordinate clause: 573 = 36% 57 = 43.5% 23 = 48% 5 = 26% 28 = 21%

relative 263 = 46% 29 = 52%

(including 2 with existential sum + dat.; 2 with pluperfect; 1 praedicabant + ind.

statement)

13 = 54% 1 = 20% 16 = 57%

(including 1 with existential sum + dative)

(21)

(relative) noun 2 (including 1 volebat + acc. + inf.)

1 1 = 20% 0

quia 0 0 3 = 12.5% 0 2

causal quod 193 = 34% 16 = 29%

(including 1 with pluperfect and 1 volebat + acc. + inf.)

3 = 12.5% 0 0

Adverbial ut 28 = 5% 0 2 = 8% 0 0

etsi 24 = 4% 1 (videbat

+ ind.

statement)

0 0 0

dum 1 0 ⁰ ⁰ ^{5 = 18%}

quā (where); ubi (where);

unde 6 = 1%; 5; 1 2 quā 0 0 3 quā;

1 unde

cum (+ indic.) 3 1 ⁰ ^{2 = 40%} ¹

postquam 4 0 0 1 = 20% 0

Other 45 6 (3 in quod

noun clauses [2 with ind.

statement]; 2 comparative clauses; and 1 condition)

1 0 0

Table 6: Contexts of Imperfect Indicative Verbs

a. This use is characteristic of Vergil. Cf. Williams (162 note ad 36-7).

b. For the main clauses, some have more than one characteristic context, so they may add up to more than the total. The “other” category includes clauses coordinate with or sentences continuing on after a clause with a characteristic context(s), and some iteratives/habituals, which tend to occur in bunches, as shown below, and miscellaneous others (e.g. with ideo, itaque, quidem).

An important thing to note about the imperfect tense is that it is regularly used in contrast with the simple past, often providing background information (cf.

Andersen 89-90, Bardovi-Harlig Tense 311-16). Thus, imperfects regularly occur after or with a perfect (simple past) verb (or narrative historic present). When an imperfect verb is used in a subordinate clause, the most common tense for the main verb is the perfect in Caesar, Cicero, and Vergil;¹¹ for Plautus, the main verb is pres-

11 Caesar: 30 (+1 present/perfect) vs. 12 imperfect main verbs; Cicero: 9 (+ 5 meminī) vs. 2 imperfect; Vergil: 10 vs. 5 imperfect.

(22)

ent tense for three and perfect for two. The imperfect occurs more often in the main clause than subordinate clause in Plautus and Vergil, probably reflecting in part the poetic genre, and slightly more often in Caesar; it is about evenly split between main and subordinate clauses in Cicero. Even when used in a main clause, there is usually a connection indicated with the preceding (or following), as discussed below.

Use of the Imperfect in (Background) Relative Clauses:

When the imperfect is found in a subordinate clause, they are most commonly (or tied as most common in Plautus) in a relative clause, usually providing background information, as indicated with imperfects in general. The kinds of imperfects reflect the general frequencies of use in each author. In Caesar, twenty-five of the twenty-nine (86%) relative clauses with the imperfect have stative verbs. In Cicero, ten of the thirteen (77%) relative clause verbs are stative; the remaining three are iterative/habitual. In Vergil, they are more evenly divided regular imperfect, progressives with verbs of activity, and iteratives/habitual. For example:

• Item Allobroges, qui trans Rhodanum vicos possessionesque habebant, fuga se ad Caesarem recipiunt et demonstrant sibi praeter agri solum nihil esse reliqui. (Caesar BG 1.11.4) (Note the historic present main verbs.)

• tum est Cato locutus, quo erat nemo fere senior temporibus illis, nemo prudentior; (Cicero de Am. 1.5) (Note the perfect tense main verb.)

• Constitit hīc, arcumque manu celerisque sagittas

corripuit, fidus quae tela gerebat Achates; (Vergil Aen. 1.187-88) (Note the perfect main verbs)

Background Causal/Explanatory Use of the Imperfect:

The imperfect is also often used to give a background cause or explana- tion, often with a stative, giving a background state of affairs as it were. According to Oldsjö, the imperfect is the most common tense for expressing cause in Caesar (Oldsjö 336). Caesar is unusual in not using the conjunction quia, but only quod, which, I believe, has influenced textbooks and led to overextended and uncharacter- istic uses.¹² Besides quia and quod subordinate clauses, the cause is often marked

12 For example, Wheelock replaced an original quia with quod in Caput XI, Sent. Ant. 4 (Pliny), and Thirty-Eight Latin Stories (Groton and May) has quod for an original quia in #36, line 1 (Petronius) and #40, line 22 (Livy). In my research, quod in general is more limited and is used especially with verbs of emotion (cf. OLD s.v.); its causal use is almost always after the main clause, except in Caesar, where it may interrupt the main clause. Sentence initial uses are not characteristic of Latin authors and can lead to inappropriate default expectations. There is not room for a full discussion here. Cf. Fugier and Bokelstein.

(23)

in main clauses by Nam(que) or enim. Following his general use, Vergil has more in the main clause, with mostly stative verbs and progressive verbs of activity. For example:

• Sed ut in Catone Maiore, qui est scriptus ad te de senectute, Catonem induxi senem disputantem, quia nulla videbatur aptior persona . . . (Cicero De Am. 1.4)

• Caesar, quod neque conloquium interposita causa tolli volebat neque salutem suam Gallorum equitatui committere audebat, commodissimum esse statuit omnibus equis Gallis equitibus detractis eo legionarios milites legionis decimae, cui quam maxime confidebat, imponere, ut praesidium quam amicissimum, si quid opus facto esset, haberet. (Caesar BG 1.42.6) (Note volebat + acc.

+ inf.)

• Caesar loquendi finem facit seque ad suos recepit suisque imperavit ne quod omnino telum in hostes reicerent. Nam etsi sine ullo periculo legionis delectae cum equitatu proelium fore videbat, tamen committendum non putabat . . . (Caesar BG 1.46.2-3) (Note also the imperfect in the background etsi clause and the indirect statement with the imperfects in both clauses.)

• Dissimulant, et nube cava speculantur amicti, quae fortuna viris, classem quo litore linquant, quid veniant; cunctis nam lecti navibus ibant, orantes veniam, et templum clamore petebant.

(Vergil Aen. 1.516-19)

• tum breviter Barcen nutricem adfata Sychaei, namque suam patria antiqua cinis ater habebat:

Other Connections of Imperfect with the Main Narrative:

The imperfect may also show a connection with the main narrative with a temporal adverb or clause, a connecting place indicator, an adversative or conces- sive (cf. etsi above), or an adverbial ut clause. These have a similar distribution of

(24)

stative imperfect, progressive, and habitual/iterative uses. See below for temporal adverbs with progressives. Vergil frequently uses a summary talis, tantus, etc. For example:

• Latonae tacitum pertemptant gaudia pectus:

talis erat Dido, talem se laeta ferebat

per medios, instans operi regnisque futuris. (Vergil Aen. 1.502-04)

• “. . . demoror, ex quo me divum pater atque hominum rex fulminis adflavit ventis et contigit igni.”

Talia perstabat memorans fixusque manebat.

• Tantos illa suo rumpebat pectore questus: (Vergil Aen. 4.553)

• obstipuit retroque pedem cum voce repressit.

improvisum aspris veluti qui sentibus anguem pressit humi nitens trepidusque repente refugit attollentem iras et caerula colla tumentem, haud secus Androgeos visu tremefactus abibat.

• hīc Hecuba et natae nequiquam altaria circum, praecipites atra ceu tempestate columbae,

condensae et divum amplexae simulacra sedebant.

(Vergil Aen. 2.515-17) Existential Use of the Imperfect:

One use of the imperfect in a main clause in Caesar, Cicero, and Vergil is that of the existential (“there was” instead of the linking) use of sum at or near the beginning of a sentence; there was no such use in Plautus’ Mostellaria. When it con- nects with the preceding and does not start a new paragraph, the existential erat is usually in its “regular” initial position. However, in my analysis it occurs more often at the beginning of a new paragraph and sets up an extended background description or temporal setting for a following action, again often with perfect tense verbs. The paragraph break is marked by the lack of a connecting adverb or particle and the initial new subject, not just a change to a different, but to a previously unknown one

(25)

– “a ___” vs. “the _____.” Note that in these the existential sum is thus moved out of initial position and is often between the subject and an adjective that is not a comple- ment, in my observations. In poetry, the non-initial order is also metri causa. When it is plural, however, in my analysis it often sets up following sub-group(s), with the existential verb in its “regular” initial position and the subject after.¹³ For example:

• Erat inter Labienum atque hostem . . . flumen . . . (Caesar BG VI.7.5 continuation of the paragraph with Labienus the subject of the previous sentence and hostes also old information)

• Planities erat magna et in ea tumulus terrenus satis grandis. . . . Eo, ut erat dictum, ad colloquium venerunt. (Caesar BG 1.43.1-2 beginning a new paragraph; note the position of the attributive adjective)

• Tempus erat, quo prima quies mortalibus aegris incipit, et dono divum gratissima serpit.

In somnis ecce ante oculos maestissimus Hector visus [est] adesse mihi largosque effundere fletus, (Vergil Aen. 2.268-71 new paragraph)

• Erant omnino itinera duo quibus . . . unum . . . alterum . . . (Caesar BG 1.6.1-2 beginning a new paragraph, but note the initial plural erant and following subject setting up the following sub-groups) Caesar: Imperfect Verbs of Thinking:

As indicated above, mental verbs of thinking (with indirect statement) are among the most frequently used verbs in the imperfect in Caesar. Joseph shows how Caesar contrasts imperfect verbs of thinking with perfect tense verbs in BG I.7 by having a paragraph of perfect tense action verbs and beginning of the next paragraph with historic present tense verbs followed by multiple verbs of thinking in the imperfect to slow the pace and thus express the careful, deliberate, and thorough thought process of Caesar (155-56). Compare a similar use also above under the Background Causal/Explanatory Use of the Imperfect.

13 Cf. new paragraph initial plural present sunt in Vergil Aen. 6. 893-95 (sunt geminae somni portae, quarum altera . . . altera . . .). I did not find any examples with the plural imperfect erant, which could not be initial metri causa, in Vergil.

(26)

Progressive Use of the Imperfect:

Besides in other contexts characteristic of imperfects in general, as above, progressives also occur in several specific characteristic contexts. When an imperfect is joined with a preceding pluperfect in the same clause (or a coordinate) clause, it is often a progressive, especially with iam.¹⁴ For example:

• inde in Allobrogum fines, ab Allobrogibus in Segusiavos exercitum ducit. Hi sunt extra provinciam trans Rhodanum primi. Helvetii iam per angustias et fines Sequanorum suas copias traduxerant et in Haeduorum fines pervenerant eorumque agros populabantur.

(Caesar BG 1.10.4-11.1)

• Ter circum Iliacos raptaverat Hectora muros, exanimumque auro corpus vendebat Achilles.

In general, iam(que) or other temporal adverbs, such as intereā, may anticipate a progressive use for an activity verb or, less often, a verb of accomplishment or achievement, following the typical lexical matching. These often indicate a shift to a more vivid narrative. For example:

• Corripuēre viam interea, quā semita monstrat.

Iamque ascendebant collem, . . .

(Vergil Aen. 1.418-19; note the perfect tense before)

• . . . huic cervixque comaeque trahuntur per terram, et versā pulvis inscribitur hastā.

Interea ad templum non aequae Palladis ibant

crinibus Iliades passis peplumque ferebant (Vergil Aen. 1.477-80) Vergil regularly uses ecce with a progressive in a main clause. For example:

• . . . Stetit illa tremens, uteroque recusso, insonuēre cavae gemitumque dedēre cavernae;

. . .

Ecce, manus iuvenem interea post terga revinctum pastores magno ad regem clamore trahebant Dardanidae, . . . (Vergil Aen. 2. 52-53, 57-59)

14 Note that in Caesar BG 1.40.15, the et . . . et marks the imperfect as a non-progressive use.